🎟 Welcome to the Token Club · GLM 5.2 now live

Open-source AI, served at the speed of light

One OpenAI-compatible endpoint for every open model worth running — GLM 5.2, Llama 4, DeepSeek, Qwen and more. Deployed on our own GPUs at the best price per token on the market, with a 90% discount on every cached token.

Get API key View models
99.9%
Uptime SLA
<100ms
Time to first token
140+
Tokens / sec
90% off
Cached tokens
~ chat.completion 
# Point the OpenAI SDK at tokentoken — that's the whole migration.
from openai import OpenAI

client = OpenAI(
    base_url="https://api.tokentoken.club/v1",
    api_key="tk-••••••••",
)

resp = client.chat.completions.create(
    model="glm-5.2",
    messages=[{"role": "user", "content": "ship it"}],
)
The model catalog

Frontier open models, one base URL

Every model is deployed, quantized and load-balanced by us — flagship quality at the best price per token on the market, and 90% off every cached token. No GPU babysitting.

GLM 5.2
355B · MoE

Our flagship reasoning + agentic model. Tool-calling and long-context tuned for production agents.

$0.15 / $0.55
in · out per M
cached in $0.015 · 90% off
142 t/sthroughput
Llama 4 Maverick
400B · MoE

Meta's multimodal workhorse. Massive context window, rock-solid instruction following at scale.

$0.20 / $0.60
in · out per M
cached in $0.020 · 90% off
128 t/sthroughput
DeepSeek V3
671B · MoE

Elite coding and math at a value price. The go-to for high-volume codegen pipelines.

$0.27 / $1.10
in · out per M
cached in $0.027 · 90% off
96 t/sthroughput
Qwen 3
235B · MoE

Blazing multilingual generalist with a hybrid thinking mode. Great quality-to-cost ratio.

$0.18 / $0.54
in · out per M
cached in $0.018 · 90% off
156 t/sthroughput
Mixtral 8x22B
141B · MoE

Mistral's sparse mixture-of-experts. Fast, cheap, and dependable for everyday chat + RAG.

$0.12 / $0.48
in · out per M
cached in $0.012 · 90% off
168 t/sthroughput
DeepSeek R1
671B · reasoning

Deep chain-of-thought reasoning that rivals closed frontier models — for complex, multi-step tasks.

$0.55 / $2.19
in · out per M
cached in $0.055 · 90% off
84 t/sthroughput
Drop-in replacement

Change two lines. Keep your whole stack.

tokentoken speaks the OpenAI API you already use. Swap the base URL and key — every SDK, framework and tool just works. No rewrites, no vendor lock-in.

  • Same schema — chat, completions, embeddings, streaming, function calling.
  • Works everywhere — LangChain, LlamaIndex, Vercel AI SDK, the raw HTTP call.
  • Sub-100ms TTFT — global GPU regions with automatic failover. 
curl https://api.tokentoken.club/v1/chat/completions \
  -H "Authorization: Bearer $TOKENTOKEN_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.2",
    "messages": [
      {"role": "user", "content": "Explain MoE in one line."}
    ],
    "stream": true
  }'
from openai import OpenAI

client = OpenAI(
    base_url="https://api.tokentoken.club/v1",
    api_key=os.environ["TOKENTOKEN_KEY"],
)

stream = client.chat.completions.create(
    model="deepseek-v3",
    messages=[{"role": "user", "content": "Explain MoE."}],
    stream=True,
)
for chunk in stream:
    print(chunk.choices[0].delta.content or "", end="")
import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api.tokentoken.club/v1",
  apiKey: process.env.TOKENTOKEN_KEY,
});

const res = await client.chat.completions.create({
  model: "qwen-3",
  messages: [{ role: "user", content: "Explain MoE." }],
});

console.log(res.choices[0].message.content);
Pay per token · No minimums · No membership fee

Welcome to the Token Club

Grab a key, keep your OpenAI code, and get $5 in free credits to benchmark it yourself. Best price per token on the market, 90% off cached — no cluster to rent, no cover charge. Members build in under a minute.

Join the club Read the docs